Wednesday, January 30, 2008

Must Employees Consent?

Two recent items concerning IRBs and studies of job performance caught my eye.

The more prominent item was a December 30, 2007, New York Times op-ed, "A Lifesaving Checklist" by Dr. Atul Gawande. Gawande describes efforts by the Johns Hopkins University and the Michigan Health and Hospital Association to offer five-step checklists to intensive-care providers, to help them remember to wash their hands and take other steps to avoid infection. OHRP shut the project down, even though it had been ruled exempt by the Hopkins IRB. (So much for local control.)

Although I try to keep this blog focused on issues concerning the humanities and social sciences, and to avoid issues concerning strictly medical research, I was struck by OHRP's determination that "the subjects of the research were both the healthcare providers at the participating ICUs and their patients." Are healthcare providers really human subjects in a case like this?

A similar question arose in a non-medical setting a few weeks later. In her pseudonymous blog on Education Week, eduwonkette complained that New York City teachers were being studied without their consent and without the approval of an IRB. Since eduwonkette herself concedes that "teachers did not need to consent in this case, as they are government employees and their employers can collect whatever data they want," I don't understand why she thought an IRB should be involved. But it does suggest that the question of studying employee effectiveness is not limited to medical contexts.

Poking around in my research notes, I find that at the July 1977 meeting of the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, Dr. Karen Lebacqz raised just this issue as a hypothetical:

"Suppose you want to do a study that involves changing the delivery of services in a mental institution ward. The patients in the institution are the ones that you want to study in terms of what the impact is on their recovery rates and so on. The staff who work on that ward will also be very directly impacted by the research. Do they constitute subjects, even though the design and the purpose of the research is not set up to study the impact on them of what you are going to do?"

While Albert Jonsen initially suggested that the staff would constitute subjects, other commissioners disagreed:

Lebacqz. “There is a very direct impact on the licensed staff.

Ryan. “Then the administrator can force it on them as a condition of employment.

Turtle. “Yes, I was going to say, that is the whole key, it is a condition of employment. Many things that, while I am a great libertarian here, lots of things I ask the people who work for me to do, I don’t ask them for their informed consent.”

[National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, Transcript, Meeting #32, July 1977 (Bethesda: The Commission, 1977), 133-135.]

After this exchange, the commission moved onto another subject, suggesting a consensus that employees are not human subjects whose informed consent is required. In its recent determination letters on the checklist project, OHRP has ruled the opposite.

I don't expect the compliance officers at OHRP to be familiar with the commission's deliberations, and I fault the commission for not defining human subjects more carefully. But it's worth noting how far OHRP has strayed from the commission whose work nominally informs its decisions.


Anonymous said...

I disagree with your conclusion. Both the Belmont Report and the Common Rule are very clear that subjects of research must give informed consent in all but a very narrow range of circumstances. Being an employee is not an exception. OHRP is simply following the legal standard set out in the Common Rule. The problem is not its interpretation, but a massive flaw in the Common Rule itself. There are many kinds of research where informed consent should not be required, and it should not be mandated or even presumptively preferred in those cases. This is one of the obvious categories where the rule is wrong and needs to be changed.

Zachary M. Schrag said...

Thanks for this comment.

I cannot agree that the Belmont Report is "very clear" on this issue, particularly since it does not conform to the sentiments the commissioners expressed in their deliberation. And rather than specifying "a very narrow range of circumstances" in which consent is not needed, the Belmont Report confesses ignorance about a vast and undefined realm called "social experimentation."

Zachary M. Schrag said...

Also, please see Tom Beauchamp's comments on this matter:

Because the definition is so nonspecific, regulatory requirements that use the definition may judge that some activities that are questionably research involving human subjects nonetheless must be treated as such. Government requirements are today commonly applied even if 'human subjects' may not need to be protected by the rules of human-subjects research. A sweeping – that is, all-inclusive conception – of 'human-subjects research' can have immediate and unjustifiable practical impact on attempts to upgrade medical care . . ."

Beauchamp Derides Federal Definition of Research

Anonymous said...

I won't disagree with you on the Belmont Report, though the document itself does not acknowledge the limitations you point out. The essential point I was trying to make is that the Common Rule is unambiguous that informed consent is required if there is more than minimal risk to the subjects. Hence, any research on employees that could lead to their firing is presumptively illegal under the rule. Given the almost unbounded scope of activities that count as "research" under the Common Rule, OHRP was essentially fulfilling its bureaucratic mission. The problems (plural) lie with the Rule, not with with some major misinterpretation by bureaucrats. One could certainly argue that in this case the employees weren't really the "subjects" of the research, or that it wasn't really "research" and I don't disagree with those arguments. But if the employees were research subjects, then voluntary informed consent was required. That is a crazy result, and one created by the text of the Rule.

Zachary M. Schrag said...

Thanks for this comment.

I should have said earlier that I agreed with you about the "massive flaw in the Common Rule itself," and only disagreed about your reading of the Belmont Report. Both documents cause headaches, and those headaches are aggravated the National Commission's failure to reconcile the Belmont Report with its IRB recommendations, which formed the basis of the Common Rule.

The ANPRM represents an overdue acknowledgment of many of the flaws in the Common Rule, but it fails to ask about the Belmont Report. In my personal comments on the ANPRM, I suggested that "the Belmont Report should be retired and replaced with a statement on research ethics that can be updated to reflect current thinking and experience."